give `OnlineLuminosityRecord` info to HLT's `LumiMonitor` plugin #39859

missirol · 2022-10-26T14:27:22Z

resolves #39756

PR description:

This PR address the 2nd item in the description of #39756, plus a related comment in #39756 (comment).

The LumiMonitor DQM plugin, used inside HLT, is updated to read the values of online instantaneous luminosity and mean PU from the so-called 'online-metadata' digis, when the corresponding info from SCAL is not available (i.e. from Run 3 onwards).

To improve clarity, some of the plot labels, transient variables, and configuration parameters are renamed. I changed "scal lumi/PU" to a more generic "online lumi/PU", but of course feel free to suggest better names.

Minor technical improvements (e.g. const-ness of class data members) were done in the process.

375912f fixes a couple of pixel-related harvesting outputs which depend on the outputs of the LumiMonitor plugin; please see #39859 (comment) for details.

(In case this is backported to 12_4_X, it should not warrant a new ConfDB parsing, since the new configuration parameter of LumiMonitor, i.e. onlineMetaDataDigis, already defaults to the intended collection; at the same time, hltOnlineMetaDataDigis should in principle be included in the DQMHistograms EndPath in the HLT menus.)

PR validation:

A local test confirms that the plots in question (profile vs PU, vs lumi) now have entries (see ).

If this PR is a backport, please specify the original PR and why you need to backport that PR. If this PR will be backported, please specify to which release cycle the backport is meant for:

CMSSW_12_4_X
CMSSW_12_5_X

missirol

Some comments inline.

missirol · 2022-10-26T14:29:19Z

DQM/HLTEvF/plugins/LumiMonitor.cc

+    online_pu = scalit->pileup();
+  } else if (onlineMetaDataDigisHandle.isValid()) {
+    online_lumi = onlineMetaDataDigisHandle->instLumi();
+    online_pu = onlineMetaDataDigisHandle->avgPileUp();
  }


The logic of this choice was inspired by

cmssw/DQM/TrackerCommon/plugins/DetectorStateFilter.cc

Line 151 in d03a9e7

if (dcsStatus.isValid() && !dcsStatus->empty()) {

and then I noticed it is basically the same as

cmssw/CalibTracker/SiStripCommon/plugins/ShallowEventDataProducer.cc

Line 67 in 67d9022

if (lumiScalers.isValid() && !lumiScalers->empty()) {

missirol · 2022-10-26T14:30:20Z

DQM/Integration/python/clients/hlt_dqm_clientPB-live_cfg.py

 process.fastTimerServiceClient.doPlotsVsPixelLumi = False
-process.fastTimerServiceClient.scalLumiME = cms.PSet(
+process.fastTimerServiceClient.onlineLumiME = cms.PSet(
    folder = cms.string('HLT/LumiMonitoring'),


This is the only place I could find where the parameter names appeared, and needed to be changed.

missirol · 2022-10-26T14:36:02Z

DQM/HLTEvF/plugins/LumiMonitor.cc

+      trkTopoToken_(doPixelLumi_ ? esConsumes<TrackerTopology, TrackerTopologyRcd>()
+                                 : edm::ESGetToken<TrackerTopology, TrackerTopologyRcd>()),


This is meant to avoid consuming the record when the plugin does not make use of it.

The following would be slightly more compact

Suggested change

trkTopoToken_(doPixelLumi_ ? esConsumes<TrackerTopology, TrackerTopologyRcd>()

: edm::ESGetToken<TrackerTopology, TrackerTopologyRcd>()),

trkTopoToken_(doPixelLumi_ ? esConsumes<TrackerTopology, TrackerTopologyRcd>() : trkTopoToken_),

but I haven't seen this approach elsewhere, and I don't know if this is a good idea.

The even-more-compact version

Suggested change

trkTopoToken_(doPixelLumi_ ? esConsumes<TrackerTopology, TrackerTopologyRcd>()

: edm::ESGetToken<TrackerTopology, TrackerTopologyRcd>()),

trkTopoToken_(doPixelLumi_ ? esConsumes() : trkTopoToken_),

unfortunately didn't compile:

error: operands to '?:' have different types 'edm::EDConsumerBaseESAdaptor<edm::Transition::Event>' and 'const edm::ESGetToken<TrackerTopology, TrackerTopologyRcd>' 91 | trkTopoToken_(doPixelLumi_ ? esConsumes() | ~~~~~~~~~~~~~^~~~~~~~~~~~~~ 92 | : trkTopoToken_), | ~~~~~~~~~~~~~~~

As an aside: if there is feedback on this technical point, I'd be interested to know (doing x_{bool ? a : x_} in the initialiser list, and the fact that one of the options doesn't compile, which is probably expected, but I'm wondering if there is a way to make it work).

This does not seem like a good idea, because x_ would end up uninitialised ?

Simple example using integers and strings:

#include <iostream> #include <string> struct Data { public: Data(int x, std::string s) : x_{x}, s_{std::move(s)} {} Data() : x_{x_}, s_{s_} {} int x() const { return x_; } std::string const& s() const { return s_; } private: int x_; std::string s_; }; int main(void) { Data value{42, "answer"}; std::cout << value.s() << " is " << value.x() << std::endl; Data data; std::cout << data.s() << " is " << data.x() << std::endl; return 0; }

The compiler warns that x_ is initialised:

$ g++ -std=c++17 -O3 -flto -g -Wall -fPIC -MMD -march=native -mtune=native -c test.cc -o test.cc.o test.cc: In member function ‘Data::Data()’: test.cc:9:15: warning: ‘this_6(D)->x_’ is used uninitialized [-Wuninitialized] 9 | Data() : x_{x_}, s_{s_} | ^~

Worse, there is no warning about s_ being uninitialised, but I get a crash:

$ ./test answer is 42 terminate called after throwing an instance of 'std::bad_alloc' what(): std::bad_alloc Aborted (core dumped)

Indeed.. thanks the example, it's clearer now!

cmsbuild · 2022-10-26T14:36:52Z

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-39859/32758

This PR adds an extra 20KB to repository

cmsbuild · 2022-10-26T14:37:20Z

A new Pull Request was created by @missirol (Marino Missiroli) for master.

It involves the following packages:

DQM/HLTEvF (dqm, hlt)
DQM/Integration (dqm)
HLTrigger/Timer (hlt)

@Martin-Grunewald, @emanueleusai, @ahmad3213, @cmsbuild, @missirol, @jfernan2, @syuvivida, @pmandrik, @micsucmed, @rvenditti can you please review it and eventually sign? Thanks.
@batinkov, @battibass, @silviodonato, @mtosi, @Martin-Grunewald, @fwyzard, @threus, @francescobrivio this is something you requested to watch as well.
@perrotta, @dpiparo, @rappoccio you are the release manager for this.

cms-bot commands are listed here

fwyzard · 2022-10-26T14:43:52Z

DQM/HLTEvF/plugins/LumiMonitor.cc

  histograms.lumiVsLS = me;

  me = booker.bookProfile("puVsLS",
-                          "scal PU vs LS",
+                          "online PU vs LS",


could you also change PU to pileup and LS to lumisection in the labels ?

fwyzard · 2022-10-26T14:45:53Z

DQM/HLTEvF/plugins/LumiMonitor.cc

@@ -267,6 +262,7 @@ void LumiMonitor::fillDescriptions(edm::ConfigurationDescriptions& descriptions)
  edm::ParameterSetDescription desc;
  desc.add<edm::InputTag>("pixelClusters", edm::InputTag("hltSiPixelClusters"));
  desc.add<edm::InputTag>("scalers", edm::InputTag("hltScalersRawToDigi"));
+  desc.add<edm::InputTag>("onlineMetaDataDigis", edm::InputTag("hltOnlineMetaDataDigis"));
  desc.add<std::string>("FolderName", "HLT/LumiMonitoring");


could you change this to folderName, for consistency with all the other parameters that are lowercase ?

Sure. For the 12_4_X backport, we might want to avoid this imho, to spare FOG from template migrations just because of this.

OK.

How are the changes to the client configuration propagated to the online DQM ?

That's a good question. So far, I only found one place, #39859 (comment). I'll double-check, and hopefully DQM reviewers will know better than me.

fwyzard · 2022-10-26T14:46:23Z

HLTrigger/Timer/plugins/FastTimerServiceClient.cc

-  if (doPlotsVsScalLumi_)
-    fillPlotsVsLumi(booker, getter, current_path, "VsScalLumi", scalLumiMEPSet_);
+  if (doPlotsVsOnlineLumi_)
+    fillPlotsVsLumi(booker, getter, current_path, "vs_lumi", onlineLumiMEPSet_);


Suggested change

fillPlotsVsLumi(booker, getter, current_path, "vs_lumi", onlineLumiMEPSet_);

fillPlotsVsLumi(booker, getter, current_path, "_vs_lumi", onlineLumiMEPSet_);

fwyzard · 2022-10-26T14:46:32Z

HLTrigger/Timer/plugins/FastTimerServiceClient.cc

  if (doPlotsVsPixelLumi_)
-    fillPlotsVsLumi(booker, getter, current_path, "VsPixelLumi", pixelLumiMEPSet_);
+    fillPlotsVsLumi(booker, getter, current_path, "vs_pixelLumi", pixelLumiMEPSet_);


Suggested change

fillPlotsVsLumi(booker, getter, current_path, "vs_pixelLumi", pixelLumiMEPSet_);

fillPlotsVsLumi(booker, getter, current_path, "_vs_pixelLumi", pixelLumiMEPSet_);

fwyzard · 2022-10-26T14:46:44Z

HLTrigger/Timer/plugins/FastTimerServiceClient.cc

  if (doPlotsVsPU_)
-    fillPlotsVsLumi(booker, getter, current_path, "VsPU", puMEPSet_);
+    fillPlotsVsLumi(booker, getter, current_path, "vs_pileup", puMEPSet_);


Suggested change

fillPlotsVsLumi(booker, getter, current_path, "vs_pileup", puMEPSet_);

fillPlotsVsLumi(booker, getter, current_path, "_vs_pileup", puMEPSet_);

fwyzard · 2022-10-26T14:47:45Z

HLTrigger/Timer/plugins/FastTimerServiceClient.cc

-    if (doPlotsVsScalLumi_)
-      fillPlotsVsLumi(booker, getter, subsubdir, "VsScalLumi", scalLumiMEPSet_);
+    if (doPlotsVsOnlineLumi_)
+      fillPlotsVsLumi(booker, getter, subsubdir, "vs_lumi", onlineLumiMEPSet_);


Suggested change

fillPlotsVsLumi(booker, getter, subsubdir, "vs_lumi", onlineLumiMEPSet_);

fillPlotsVsLumi(booker, getter, subsubdir, "_vs_lumi", onlineLumiMEPSet_);

and so on for the others

Mh.. but one underscore is added here (I also added a screenshot to the description). Or do we really want the double underscore?

ah... no, we definitely do not want the double underscore :-)

cmsbuild · 2022-10-26T16:30:49Z

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-39859/32763

This PR adds an extra 32KB to repository

cmsbuild · 2022-10-26T16:31:15Z

Pull request #39859 was updated. @Martin-Grunewald, @emanueleusai, @ahmad3213, @cmsbuild, @missirol, @jfernan2, @syuvivida, @pmandrik, @micsucmed, @rvenditti can you please check and sign again.

missirol · 2022-10-26T16:38:28Z

Tried to address the comments received so far.

Regarding #39859 (comment), I probably need feedback from @cms-sw/dqm-l2. If it's useful, before we consider merging this and backporting it, I could make a 124X version of this PR to be tested in the online DQM (we would not see the empty HLT plots getting filled, because that requires rerunning the HLT, but we would verify that things still work okay).

fwyzard · 2022-10-26T16:55:12Z

👍🏻

cmsbuild · 2022-10-26T17:29:35Z

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-39859/32765

This PR adds an extra 32KB to repository

missirol · 2022-10-27T05:53:43Z

please test

cmsbuild · 2022-10-30T15:23:40Z

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-39859/32855

This PR adds an extra 12KB to repository

cmsbuild · 2022-10-30T15:24:06Z

Pull request #39859 was updated. @Martin-Grunewald, @emanueleusai, @ahmad3213, @cmsbuild, @missirol, @jfernan2, @syuvivida, @pmandrik, @micsucmed, @rvenditti can you please check and sign again.

missirol · 2022-10-30T15:25:59Z

please test

In Run-3 data wfs, harvesting outputs vs_lumi and vs_pileup now have entries as expected: example.

The bin-by-bin DQM differences also show two TProfiles in PixelPhase1/Tracks/hitefficiency_per_scalLumi_*; both are built based on information from the TProfile lumiVsLS, which this PR touches. For the wfs checked in PR tests, these two pixel plots seem to be somewhat broken (example from wf 11634.0, example from wf 139.001). Maybe this is because they are created with xmin == xmax (yMax2 == 0). In MC (e.g. wf 11634.0), this happens because the lumiVsLS TProfile has GetMaximum == 0. In wf 139.001 (data), lumiVsLS.GetMaximum = 0.013, but hitefficiency_per_scalLumi_*.xmax == xmin == 0, because yMax2 is of type int (so 0.013 becomes a zero).

I tried to fix this in 375912f; in the MC case, the 2 profiles wrt lumi (which would always be empty) are not produced anymore.

I don't have other changes planned for this PR.

cmsbuild · 2022-10-30T18:10:06Z

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-62cf54/28647/summary.html
COMMIT: 375912f
CMSSW: CMSSW_12_6_X_2022-10-30-0000/el8_amd64_gcc10
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/39859/28647/install.sh to create a dev area with all the needed externals and cmssw changes.

Comparison Summary

Summary:

No significant changes to the logs found
Reco comparison results: 10 differences found in the comparisons
DQMHistoTests: Total files compared: 48
DQMHistoTests: Total histograms compared: 3416356
DQMHistoTests: Total failures: 97
DQMHistoTests: Total nulls: 0
DQMHistoTests: Total successes: 3416237
DQMHistoTests: Total skipped: 22
DQMHistoTests: Total Missing objects: 0
DQMHistoSizes: Histogram memory added: 55.515 KiB( 47 files compared)
DQMHistoSizes: changed ( 1000.0,... ): 0.039 KiB DQM/TimerService
DQMHistoSizes: changed ( 138.5,... ): 27.582 KiB PixelPhase1/Tracks
Checked 206 log files, 48 edm output root files, 48 DQM output files
TriggerResults: no differences found

missirol · 2022-10-30T21:16:33Z

+hlt

differences in outputs of wfs 20896.0,20900.0,21034.999,23234.0,25202.0 are likely unrelated to this PR, and have been seen elsewhere (see Non-reproducibility in JetMET/{Jet,MET}Validation histograms in phase2 workflows #39754)
other differences are expected, and consistent with the changes in this PR

missirol · 2022-11-01T06:29:44Z

@cms-sw/dqm-l2 , any comments on this PR? It would be nice to have its backport merged in 12_4_X sooner rather than later.

emanueleusai · 2022-11-02T07:26:24Z

testing the 12_4 backport at p5 now, apologies for the delay

syuvivida · 2022-11-03T20:17:43Z

+1

cmsbuild · 2022-11-03T20:18:11Z

This pull request is fully signed and it will be integrated in one of the next master IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @perrotta, @dpiparo, @rappoccio (and backports should be raised in the release meeting by the corresponding L2)

perrotta · 2022-11-04T07:36:31Z

+1

cmsbuild added this to the CMSSW_12_6_X milestone Oct 26, 2022

cmsbuild added code-checks-pending dqm-pending hlt-pending orp-pending pending-signatures tests-pending labels Oct 26, 2022

missirol commented Oct 26, 2022

View reviewed changes

cmsbuild added code-checks-approved and removed code-checks-pending labels Oct 26, 2022

missirol mentioned this pull request Oct 26, 2022

Issues related to FastTimerService and HLT #39756

Closed

fwyzard reviewed Oct 26, 2022

View reviewed changes

missirol force-pushed the devel_onlineMetaDataToLumiMon branch from 38f4785 to 4ce015f Compare October 26, 2022 16:22

cmsbuild added code-checks-pending and removed code-checks-approved labels Oct 26, 2022

cmsbuild added code-checks-approved and removed code-checks-pending labels Oct 26, 2022

cmsbuild added code-checks-pending tests-pending labels Oct 30, 2022

cmsbuild added code-checks-approved and removed code-checks-pending labels Oct 30, 2022

cmsbuild added tests-started and removed tests-pending labels Oct 30, 2022

cmsbuild added tests-approved and removed tests-started labels Oct 30, 2022

cmsbuild added hlt-approved and removed hlt-pending labels Oct 30, 2022

cmsbuild added dqm-approved fully-signed and removed dqm-pending pending-signatures labels Nov 3, 2022

cmsbuild added orp-approved and removed orp-pending labels Nov 4, 2022

cmsbuild merged commit 56f444d into cms-sw:master Nov 4, 2022

cmsbuild mentioned this pull request Nov 4, 2022

Photon MVA-based ID for Run3 from EGamma POG #39927

Merged

missirol deleted the devel_onlineMetaDataToLumiMon branch November 5, 2022 22:32

missirol mentioned this pull request Dec 15, 2022

add unit test for HLT online-DQM plugins #40334

Merged

		trkTopoToken_(doPixelLumi_ ? esConsumes<TrackerTopology, TrackerTopologyRcd>()
		: edm::ESGetToken<TrackerTopology, TrackerTopologyRcd>()),

	trkTopoToken_(doPixelLumi_ ? esConsumes<TrackerTopology, TrackerTopologyRcd>()
	: edm::ESGetToken<TrackerTopology, TrackerTopologyRcd>()),
	trkTopoToken_(doPixelLumi_ ? esConsumes() : trkTopoToken_),

	fillPlotsVsLumi(booker, getter, current_path, "vs_lumi", onlineLumiMEPSet_);
	fillPlotsVsLumi(booker, getter, current_path, "_vs_lumi", onlineLumiMEPSet_);

	fillPlotsVsLumi(booker, getter, subsubdir, "vs_lumi", onlineLumiMEPSet_);
	fillPlotsVsLumi(booker, getter, subsubdir, "_vs_lumi", onlineLumiMEPSet_);

give OnlineLuminosityRecord info to HLT's LumiMonitor plugin #39859

give OnlineLuminosityRecord info to HLT's LumiMonitor plugin #39859

Conversation

missirol commented Oct 26, 2022 • edited Loading

PR description:

PR validation:

If this PR is a backport, please specify the original PR and why you need to backport that PR. If this PR will be backported, please specify to which release cycle the backport is meant for:

missirol left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

missirol Oct 26, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

cmsbuild commented Oct 26, 2022

cmsbuild commented Oct 26, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

cmsbuild commented Oct 26, 2022

cmsbuild commented Oct 26, 2022

missirol commented Oct 26, 2022

fwyzard commented Oct 26, 2022

cmsbuild commented Oct 26, 2022

missirol commented Oct 27, 2022

cmsbuild commented Oct 30, 2022

cmsbuild commented Oct 30, 2022

missirol commented Oct 30, 2022

cmsbuild commented Oct 30, 2022

Comparison Summary

missirol commented Oct 30, 2022 • edited Loading

missirol commented Nov 1, 2022

emanueleusai commented Nov 2, 2022

syuvivida commented Nov 3, 2022

cmsbuild commented Nov 3, 2022

perrotta commented Nov 4, 2022 • edited Loading

give `OnlineLuminosityRecord` info to HLT's `LumiMonitor` plugin #39859

give `OnlineLuminosityRecord` info to HLT's `LumiMonitor` plugin #39859

missirol commented Oct 26, 2022 •

edited

Loading

missirol Oct 26, 2022 •

edited

Loading

missirol commented Oct 30, 2022 •

edited

Loading

perrotta commented Nov 4, 2022 •

edited

Loading